Skip to main content

​
Get Started

Follow these steps to instrument your first LLM application with Anyway.
1

Create an Account

Sign up at app.anyway.sh and create your first project.Once logged in, navigate to Developer > API Keys and create a new SDK API key.
Keep your API key secure. Never commit it to version control.
2

Install the SDK

Install the Anyway Python SDK using pip (requires Python 3.10+):
pip install anyway-sdk
3

Initialize the SDK

from anyway.sdk import Traceloop

Traceloop.init(
    app_name="my-app",
    api_endpoint="https://collector.anyway.sh/",
    headers={"Authorization": "your-api-key"}
)
4

Trace Your LLM Calls

The SDK automatically instruments LLM calls from OpenAI, Anthropic, and other providers.Add workflow and task structure to your traces:
from anyway.sdk.decorators import workflow, task
from openai import OpenAI

client = OpenAI()

@task(name="chat")
def chat(prompt: str) -> str:
    response = client.chat.completions.create(
        model="gpt-4o-mini",
        messages=[{"role": "user", "content": prompt}]
    )
    return response.choices[0].message.content

@workflow(name="answer_question")
def answer_question(question: str) -> str:
    return chat(question)

result = answer_question("Hello!")
5

View Your Traces

Open the Anyway Dashboard and navigate to Traces.You’ll see your workflow and task spans with:
  • Execution hierarchy (workflow → tasks)
  • Duration and timing for each operation
  • Request and response data
  • Token usage and cost estimates
Click on any trace to see the full span tree.

​
Next Steps

Python SDK

Configure the Python SDK.

JavaScript SDK

Configure the JavaScript SDK.

Tracing Concepts

Learn about distributed tracing.

Cost Tracking

Monitor your AI spend.

​
Need Help?

Contact Support

Have questions? We’re here to help.